MST 502Mastery · 90 days3 US credit hours·Mastery course · Upper-division / professional elective·v1 · updated 3 Oct 2026

Harness Engineering Mastery

University / Board Instructor Student Outcomes & Skills

Course outcomes

CO1

Demonstrate Foundations & Theory competence: complete the Day 1–15 builds, explain the underlying theory, and defend the phase artefact under live questioning.

C2 Understand

Harness architecture (model + harness)Foundation model mechanics Structured outputs & schema validation Agent environments & sandboxing Building an agent loop from primitives (no framework)

CO2

Demonstrate Core Building · the single agent competence: complete the Day 16–30 builds, explain the underlying theory, and defend the phase artefact under live questioning.

C3 Apply

Building an agent loop from primitives (no framework)Tool contract design, descriptions and token-efficient returns Context engineering RAG pipeline design Agent memory & state Tool calling & MCP integration

CO3

Demonstrate Orchestration & Integration competence: complete the Day 31–45 builds, explain the underlying theory, and defend the phase artefact under live questioning.

C4 Analyse

Graph orchestration with typed state and conditional control flow Model selection, routing, fallback and provider portability Agent memory, budgets & guardrails Running open models locally (Ollama)Multi-agent orchestration

CO4

Demonstrate Evaluation & Observability competence: complete the Day 46–60 builds, explain the underlying theory, and defend the phase artefact under live questioning.

C4 Analyse

LLMOps observability & SLOs Evaluation engineering with golden sets LLM-as-judge calibration CI regression gates for AI

CO5

Demonstrate Production, Security & Governance competence: complete the Day 61–75 builds, explain the underlying theory, and defend the phase artefact under live questioning.

C5 Evaluate

Agent environments & sandboxing Agent memory, budgets & guardrails AI security & prompt-injection defence Policy-as-code for agents (OPA, Cedar)AI red-teaming AI code review, testing & supply chain Responsible AI & regulatory compliance

CO6

Demonstrate Advanced, Autonomous & Mastery competence: complete the Day 76–90 builds, explain the underlying theory, and defend the phase artefact under live questioning.

C6 Create

Agent memory & state Multi-agent orchestration Long-running & self-verifying agents LLM cost & latency estimation Failure taxonomy, improvement flywheel, labelling operations and incident response Team delivery & technical communication

Phases & skills

Phase	Topic	Outcomes	Deliverable	Skills
1	Foundations & Theory Days 1–15 · C2	CO1	CAPSTONE (Phase 1) — the typed augmented LLM + architecture brief	Harness architecture (model + harness)Foundation model mechanics Structured outputs & schema validation Agent environments & sandboxing Building an agent loop from primitives (no framework)
2	Core Building · the single agent Days 16–30 · C3	CO2	CAPSTONE (Phase 2) — the single-agent harness	Building an agent loop from primitives (no framework)Tool contract design, descriptions and token-efficient returns Context engineering RAG pipeline design Agent memory & state Tool calling & MCP integration
3	Orchestration & Integration Days 31–45 · C4	CO3	CAPSTONE (Phase 3) — durable, model-agnostic agent with gates	Graph orchestration with typed state and conditional control flow Model selection, routing, fallback and provider portability Agent memory, budgets & guardrails Running open models locally (Ollama)Multi-agent orchestration
4	Evaluation & Observability Days 46–60 · C4	CO4	CAPSTONE (Phase 4) — the instrumented, gated harness	LLMOps observability & SLOs Evaluation engineering with golden sets LLM-as-judge calibration CI regression gates for AI
5	Production, Security & Governance Days 61–75 · C5	CO5	CAPSTONE (Phase 5) — the hardened, governed harness	Agent environments & sandboxing Agent memory, budgets & guardrails AI security & prompt-injection defence Policy-as-code for agents (OPA, Cedar)AI red-teaming AI code review, testing & supply chain Responsible AI & regulatory compliance
6	Advanced, Autonomous & Mastery Days 76–90 · C6	CO6	MASTERY CAPSTONE III — harden, benchmark, document, defend	Agent memory & state Multi-agent orchestration Long-running & self-verifying agents LLM cost & latency estimation Failure taxonomy, improvement flywheel, labelling operations and incident response Team delivery & technical communication